In the first part of the project, VERTEXSO investigated observational data from profiling floats that freely drift in the ocean and moorings that are attached to the sea floor. In particular, we investigated data from austral winter to find evidence for the existence of convective plumes in the sea-ice covered Southern Ocean. This analysis revealed a widespread occurrence of sporadic interactions between winter water that is relatively cold, fresh, and rich in dissolved oxygen and the upwelling deep water that is relatively warm, saline, and poor in dissolved oxygen. The largest variations in these properties in time and space, identified by the profiling floats, are found at the base of the winter water, which is marked by the upper limit of the so-called pycnocline, i.e. the main density gradient that separates the surface from the deep ocean. We find that this variability is associated with the upwelling and mixing of deep water into the surface layer, which mostly occurs in the seasonally ice-covered Southern Ocean. In analysis of the temporal evolution of this mixing, we find that this variability is largely induced by the ventilation of the upper pycnocline water during austral winter when the water column is only weakly stratified.
A more detailed analysis of the temporal evolution of the upper ocean water column throughout the winter is performed on a local mooring site in the Weddell Sea. This analysis reveals distinct vertical mixing events that reach the pycnocline in winter and presumably represent convective plumes in the water column. The plumes develop in late winter when the upper ocean salinity has substantially increased due to sea ice formation. They occur together with a decline in the sea ice cover and the direct linkages are still being explored. The plumes are marked by a sharp decrease in the subsurface temperature and salinity during periods of weak density stratification, when presumably cold and fresh waters are entrained into the upwelling deep water.
Another ongoing investigation attempts to reproduce these convective plumes observed on the mooring site with a non-hydrostatic ocean model. In contrast to hydrostatic ocean models, such as the ones used for global climate simulations, this model is able to directly resolve convective processes in the ocean. The model has been successfully set up with the initial conditions at the mooring site and a convective plume could be reproduced for an idealized surface forcing that shows the ventilation of the upwelling deep water and the associated small-scale circulation and mixing processes. Further work is needed to make those simulations more realistic and analyze the model output to better understand the underlying processes.
